Question

Describe the differences between two running processes versus one process running a pair of threads. Your response should include computational costs of creation, destruction and context switching. It should also include a discussion on address spaces and resource sharing

Answer 1

ANSWER:

GIVEN THAT:

To describe the differences between two running processes versus one process running a pair of threads.

i)Let us discuss about the differences between two running processes (multiprocessing) versus one process running a pair of threads (multithreading)

ii)The main difference between these two techniques is that,in multi-threading if any one of the thread goes wrong,then the entire process will goes down where as in multiprocessing the modules has different memory regions and even one process goes down others can execute.

iii)Also, in multiprocessing the message delivery is not guaranteed since the reviever may crashed before recieving the message but in multi-threading, we can assume that message delivery is guaranteed

iv)In multi-threading the entire memory is shared by whole threads so messages can be passed to other objects where as in multiprocessing we need to implement shared memory between processes.

v)Inorder to send and receive message in multi-threading objects can directly communicate ech other where as in multiprocessing techniques like socket,messge passing or shared memory has to be implemented.

vi)multi-threading allow us to use same adress space for all the threads executing where multiprocessing needs to specify adress for each process from the adress space.

So we can conclude that

vii)Comparing to multiprocessing, multithreading has the advantage of less overhead to create and terminate since only a small memory is needed to copied

viii)also multi threading enables the faster task switching since context switching is easier in threads since not the entire information about the process is to be stored for switching the process only the small thread stack is to be copied and replaced.

ix)Since threads share a common memory among all the threads the applications that are supposed to share large amout of data between objects can be easily implemented using multi-threading than multi-processing.but if any error occure in that memory space entire application will goes down.